A rank tracker is software that records where a website appears in a search engine's results for a defined list of keywords, and how that position changes over time. It exists because search engines do not hand site owners a daily, keyword-by-keyword ranking report. The tool fills the gap by running the searches itself, usually automatically, reading the results page and storing the position of each tracked domain. Agencies, in-house search teams and publishers use the resulting series to report progress, spot losses and watch competitors.

The term covers standalone products such as AccuRanker and Nozzle, modules inside suites such as Semrush's Position Tracking and Ahrefs' Rank Tracker, and data providers such as SerpApi and DataForSEO, which sell raw results that other tools turn into rankings.

How a rank tracker works

The process starts with a keyword set: the queries that matter commercially, the domain to watch and its competitors. Semrush's Position Tracking allows up to 20 competitor domains per campaign, according to its knowledge base.

Each keyword is then paired with a search context. Google personalises and localises results, so the same query returns different pages in Leeds and in Lisbon, on a phone and on a laptop. Trackers specify the country (Google's gl parameter), the interface language (hl), the device and often a city or postcode. Local precision usually relies on the UULE parameter, short for Universal URL Encoded Location, an encoded string that tells Google where the search is supposedly coming from. According to SerpApi, one version encodes a canonical place name such as "West New York, New Jersey, United States", while a second version encodes GPS coordinates.

Next comes collection. On a fixed schedule, typically daily, the tracker requests the results page. Google offers no ranking API for third parties, so almost every tracker scrapes the public page, directly or through a SERP API vendor. Requests are routed through pools of residential proxy IP addresses so the traffic resembles ordinary home users.

The page is then parsed: software identifies each organic listing, ad, map pack, carousel and AI Overview, extracts destination URLs and assigns positions. Conventions differ by vendor. Semrush, for instance, counts a Knowledge Panel as position 11 and treats a keyword that does not rank at all as position 100 when calculating average position, according to its documentation.

Finally, positions become metrics. Most tools add aggregates to the raw table of keyword, date and position: average position, keywords in the top three or top 10, and a weighted "visibility" score. Semrush's Visibility % weights positions by estimated click-through rate in the top 100, so 100% means first place for every tracked keyword. Its documentation gives an example: two sites both average position 50 across two keywords, but the one ranking first and 100th scores higher than the one ranking 50th twice, because the top spot draws most clicks.

Location changes the answer. A Manchester plumbing firm tracking "emergency plumber" may sit at position 14 in a national, unlocalised check and third on mobile from a Manchester postcode. Both numbers are correct; they describe different searchers.

Where the data differs from Google's own

The main alternative source is Google Search Console, which reports clicks, impressions and position from Google's own logs. Its definition is different. Search Console uses the "topmost position occupied by a link to your property" for each query, then averages across queries where the site actually appeared, according to Google's help documentation. A query with no impression contributes nothing. An AI Overview "occupies a single position" and every link inside it shares that position.

A rank tracker therefore shows one sampled view at a chosen time and place, for every keyword on the list, ranked or not. Search Console averages all real searches, but only where the site was shown. The two figures rarely match, and neither is wrong.

Origin and evolution

Desktop rank-checking software appeared in the late 1990s, as search engine optimisation became a trade. WebPosition Gold, made by FirstPlace Software, became the best-known product. Search engines objected early. In June 2007, Google's webmaster guidelines stated that its terms "do not allow the sending of automated queries of any sort" and that Google "does not recommend the use of products such as WebPosition Gold", according to a contemporary account by Marty Weintraub of aimClear. By then the product line belonged to NetIQ, maker of the WebTrends analytics software, which announced the acquisition in April 2004.

The tension persists. Google's current spam policies, last updated on August 28, 2026, define machine-generated traffic as automated queries to Google and specifically include "scraping results for rank-checking purposes" conducted without express permission. The industry nonetheless moved to cloud services checking thousands of keywords daily, and consolidated: Moz bought STAT Search Analytics, a 50-person Vancouver tracker, on October 10, 2018, its largest acquisition at the time, according to GeekWire.

The more recent history is one of rising cost. Until September 2025, a tracker could add &num=100 to a Google URL and receive 100 results in one request. Google removed that parameter around September 14, 2025, so collecting the same 100 positions took 10 requests. Vendors responded differently. AccuRanker moved daily tracking to the top 30, with full top-100 updates twice a month, according to its own September 19, 2025 notice. SISTRIX kept 100-position crawling for index and project keywords and stopped systematic desktop collection on September 22. Ahrefs began restoring top-100 tracking for enterprise customers on October 8, 2025.

Litigation sharpened the dispute. Google sued SerpApi on December 19, 2025 under the anti-circumvention provisions of the Digital Millennium Copyright Act (DMCA), alleging it bypassed SearchGuard, an anti-bot system launched in January 2025. The court dismissed Google's general DMCA claims on July 20, 2026.

Why it matters for marketers

Rankings are the most direct early proxy for organic performance. Agencies report them, finance teams use them in forecasts and content teams use them to choose which pages to update. A fall from third to ninth on a high-volume query can cost more traffic than a dozen gains lower down. Rankings also inform paid search, where a brand that ranks first organically for its own name may bid less on it. And trackers record competitors' positions, which Search Console cannot.

The concept is also expanding beyond Google's blue links. SerpApi's open-source AppTrail, released on September 30, 2026, tracks app-store rankings alongside mentions in Google AI Mode, AI Overviews and Bing Copilot. Several vendors now sell "AI visibility" or prompt-tracking modules that apply rank-tracking logic to chatbot answers, a category covered in PPC Land's AI visibility explainer.

Limitations and disputes

A ranking is a sample, not a fact. Results vary by location, device, time, personalisation and live experiments, yet one daily check stands in for millions of searches.

The sample can also be corrupted. Since September 2026, rank-tracking vendors have reported that Google serves altered pages to traffic it suspects is automated. DemandSphere put spurious results at about 11% of its daily checks and Nozzle measured 8.0% of pages as poisoned. Both used undisclosed methods on different keyword sets, and Google has confirmed neither figure. A poisoned page looks normal, which makes it more damaging than a visible block: it produces a confident number that no real searcher saw. Such episodes can mimic the short disappearances described in PPC Land's drop-and-pop explainer.

Cost limits depth, and vendors disagree about how many positions are worth tracking. Semrush said in September 2025 that top-10 rankings remained "fully reliable" and top-20 data useful, while positions further down should be treated as supplementary. Accounts differ even on what individual vendors did: PPC Land's coverage describes AccuRanker limiting tracking to the top 20, while AccuRanker's own notice describes daily tracking of the top 30.

Legitimacy is contested. Google's position, set out in its spam policies, is that unpermitted rank-checking violates its terms. Vendors argue they are collateral damage in a fight aimed at artificial intelligence companies that use search results for live retrieval. DemandSphere chief executive Ray Grieselhuber said "Google does not want the AI labs to use their SERPs as part of their live retrieval process". Google has not said who its measures target.

Search Console. Google's free first-party report, built from real impressions and clicks on verified sites. It cannot report competitors and shows position only where a site was seen; a rank tracker samples results for any domain.

SERP API. A data service, such as SerpApi or DataForSEO, that returns parsed search results on request. Many rank trackers are built on SERP APIs, but the API itself stores no history and computes no rankings for a client.

Visibility Index and MozCast. Both are built from rank-tracking data but answer different questions. A visibility index, such as SISTRIX's, condenses a domain's rankings across a large fixed keyword set into one score. MozCast, from Moz, measures how much Google's results moved overall on a given day. A rank tracker reports the positions of one client's own keyword list.

Recent developments

The cost of tracking rose again in 2026. Google began routing result links through google.com/goto redirects, confirmed on August 26, 2026. The encoded links cannot be decoded locally, so Nozzle founder Derek Perkins estimated that resolving every link in a five-page ranking takes 500 to 1,000 requests. SISTRIX began resolving each link individually, describing the result as "more complex and more expensive as a result, though not impossible", and SerpApi said on September 5 that it again returned direct URLs.

Google meanwhile expanded its own reporting. Generative AI performance reports reached all Search Console sites by August 31, 2026, showing impressions in AI features but no clicks or queries. The overlap between AI surfaces, blocked collection and the September 2026 spam update has made ranking data from autumn 2026 unusually hard to interpret, a point PPC Land made in its October 13 report.

Timeline

  • Late 1990s: Desktop rank-checking software, including WebPosition Gold from FirstPlace Software, comes into wide use.
  • April 2004: NetIQ, maker of WebTrends, announces it is acquiring WebPosition Gold.
  • June 2007: Google's webmaster guidelines name WebPosition Gold and say automated queries breach its terms.
  • October 10, 2018: Moz acquires STAT Search Analytics.
  • January 2025: Google launches SearchGuard.
  • September 14, 2025: Google removes the num=100 parameter; Semrush confirms the impact the next day.
  • September 19, 2025: AccuRanker moves daily tracking to the top 30.
  • September 22, 2025: SISTRIX stops systematic desktop SERP collection.
  • October 8, 2025: Ahrefs restores top-100 tracking for enterprise customers.
  • December 19, 2025: Google sues SerpApi under the DMCA.
  • July 20, 2026: Court dismisses Google's general DMCA claims against SerpApi.
  • August 26, 2026: Google confirms goto passthrough redirects in results.
  • August 28, 2026: Google's spam policies, naming rank-checking scrapers, last updated.
  • August 31, 2026: Search Console generative AI reports reach all sites.
  • Around September 11, 2026: DemandSphere reports intensified blocking and spurious results.
  • September 30, 2026: SerpApi releases AppTrail.
  • October 13, 2026: DemandSphere's 11% spurious-results figure is reported.

Summary

Who: Rank-tracking vendors such as AccuRanker, Nozzle, DemandSphere, Semrush, Ahrefs and SISTRIX, data suppliers such as SerpApi and DataForSEO, and the agencies, brands and publishers that buy their reports; Google, whose results they collect and whose terms prohibit unpermitted automated queries.

What: Software that repeatedly queries a search engine for a defined keyword list, records where tracked domains appear by location and device, and turns the positions into time series and aggregate metrics.

When: Desktop rank checkers date from the late 1990s; Google publicly objected by 2007; collection became markedly more expensive from September 14, 2025, with further restrictions through 2026.

Where: Mainly in Google's organic results, increasingly in local packs, app stores and AI answer surfaces such as AI Overviews and AI Mode.

Why: Search engines provide no third-party ranking feed, so trackers supply the keyword-level, competitor-inclusive data that search teams report and budget against, at a growing cost and with growing questions about data quality.